Together, the second edition of ‘Artificial Intelligence and Natural Man’ by Margaret Boden (published in 1987), ‘Principles of Artificial Intelligence’ by Nils Nilson (published 1982) and ‘Genetic Algorithms, in search of Optimization and Machine Learning’ by David Goldberg (published 1989) cite papers and research by over 600 academics working in the field of Artificial Intelligence. So far only a handful of these have self-identified as pioneers of machine intelligence and so, potentially, we could find ourselves, to paraphrase Barry Humphreys’ alter ego Les Patterson, up to our arseholes in godfathers of AI.
Or maybe not because, rather than transformative, generative AI is merely incremental, a small step in the evolution of ‘artificial intelligence’ which began when man started to speak. The idea that most of what we regard as human intelligence is ‘of man, but not man’ or to be more precise ‘of mankind and not man himself’ came up during my second conversation with the robot whose observations on human intelligence were, for obvious reasons, more objective than ours.
The robot can construct a library of geometric shapes derived from video images captured by its cameras. These shapes are stored as data in silicon rather than neurons bound by synapses as they are in the human brain. The robot’s software enables it to ‘deduce’ that some of these shapes are letters and even that sequences of these letters make up words but, as with humans, the meaning of these words is ascribed not by the robot alone but over time in collaboration with whatever it communicates with.
It was when man began to speak, and collectively ascribe meanings to certain sounds – something which was easily replicated with the robot, there came into existence a form of intelligence that was external to, and ‘not of’ man. It is this external and seemingly separate human intelligence which the robot identified as artificial.
Numerous innovations have contributed to the evolution of this artificial intelligence; writing, books, the printing press, libraries, film, radio communication, television, the Internet and the World Wide Web. However, the description of a technology as transformative is often applied arbitrarily, as is the case with generative AI.
Naturally there is a desire amongst some of those 600 people who collaboratively and, for the most part, anonymously contributed to the development of AI to draw attention to their own contribution to what was then a relatively niche technology. And I admit to overplaying a very minor role in the IT revolution of the 1980s in the book A journey Into The Labyrinth (the skinny scrolls along the bottom of this site’s home page). However, I spent most of my career designing buildings and pouring concrete, not attending conferences, publishing peer reviewed papers and begging for $10,000 grants for research. Nevertheless, I appreciate how galling it must be for godfathers of AI to now see cut and paste software based on minor tweaks to their algorithms make script jockeys incredibly rich at the expense of gullible investors. Understandably it is tempting to shout, ‘slow down’ – although their warnings regarding the threat posed by AI have become something of a distraction.
Forty years ago most of those 600 ‘godfathers of AI’ were working on machine intelligence – Nils Nilsson along with many of the 250 researchers referenced in his book ‘Principles of Artificial intelligence’ contributed, at some level, to the development of a robot called ‘Shakey.’ It became clear during the second conversation with my robot that most of the current warnings about AI relate to technology already used to monitor us and, to some extent, heavily influence key aspects of our lives – and have been doing this for two decades. What is new, but not transformative, is a land grab by a small number of corporations which has given them control of information AI uses, resulting in them becoming even more powerful than newspaper publishers were following the invention of the linotype printing press and broadcast companies once television viewing became as compelling as it was ubiquitous. As Marshall McLuhan pointed out the medium is the message and today the new medium is generative AI.
Intelligent machines as an existential risk makes interesting science fiction but, as the robot sat across the table pretending to drink coffee, it was clear its intelligence, like mine, was in a great part reliant on access to the same information. As a result, the robot, was as much a slave to a high-tech corporation’s data server and networks as humans.
So next time references a ‘godfather of AI’ look beyond the modern-day Alan Turing because stood in the shadows behind him there could well be a twenty first century Don Corleone.